OPTIMAL WAVELENGTH RANGE SELECTION BY A GENETIC ALGORITHM FOR DISCRIMINATION PURPOSES IN SPECTROSCOPIC INFRARED IMAGING

When spectroscopic infrared imaging is applied to discriminate between different materials, multiple images have to be measured at different wavelengths or wavelength ranges. The time-consuming step in present on-line spectroscopic imaging is the measurement and processing time p er identification of a number of spectroscopic images. PT this number of images can be kept small, whereby are optimal discrimination is sti ll guaranteed, the acquisition and processing time will be taster and, therefore, this approach becomes attractive: in real-world applicatio ns, This paper describes the search for a limited number of spectrosco pic wavelengths or wavelength ranges far images where optimal discrimi nation between the materials is guaranteed, This optimization is appli ed in particular to the discrimination between plastics and nonplastic s, Because the number of potential wavelength combinations is huge, a genetic algorithm (GA) is used as a subset selection technique to solv e this large-scale optimization problem. Since the problem concerns cl assification, a specific optimization criterion is developed. Finally, infrared images art? measured at the calculated optimal wavelength ra nges, and the resulting discrimination performance is compared with th at of images measured at wavelengths chosen on the basis of a priori s pectroscopic knowledge.